1. Field of the Invention
This invention relates generally to combined ballasts and wiring harnesses for fluorescent-lamp fixtures; and more particularly to so-called "leadless" ballasts that directly carry connectors for attachment to wiring in the fixtures.
2. Prior Art
Fluorescent lamps require relatively high starting voltages, and in many cases electrode heating. These are supplied by a combination of transformer coils, capacitors and thermal-overload circuit breakers, all usually potted together in a metallic enclosure familiarly known as a "ballast".
Some so-called "electronic ballasts" have much smaller, lighter coils and relatively much more extensive electronic circuitry. These units may be potted, or their components may be coated only lightly ("dipped") or not at all.
A typical indoor fluorescent-lamp fixture or luminaire is an elongated, narrow structure with an even narrower, shallow casing that extends the length of the fixture for mounting of fluorescent-lamp sockets and for housing of the ballast and the fixture wiring. As the ballast usually fits within (or sometimes upon) one of these narrow, shallow casings, the ballast too is usually made relatively long, narrow and shallow.
The ballast has its own enclosure, usually made of two sheet-metal pieces. One piece is die-cut and then bent to provide two generally vertical side walls, a generally horizontal floor, and conventionally a vertical wall at each end of the enclosure respectively. A second, flat piece (with mounting holes for attachment to the casing) forms a separate coverplate.
In this document we shall refer to the ballast by the nomenclature just established--in which the flat coverplate is considered to be the top of the ballast, and the horizontal panel that is made integrally with the side and end walls is considered to be the bottom. Ballasts are in that orientation when potting material is poured into the cans for potting the components, and usually or at least often are also mounted in that way. In any event we shall use this terminology for purposes of definiteness--although, for descriptive purposes, in many patents and other documents ballasts are shown inverted with respect to the convention just described; and ours too can be so oriented in use.
General practice in the fluorescent-lighting industry for more than a half century has been to provide wires that extend from within the ballast through a grommet or strain relief in each end wall, respectively. Some of these wires connect with a lamp socket mounted at each end of the lamp fixture, respectively; and others of the wires connect with the input power leads.
The ballast wires sometimes are made the correct length to just reach the sockets in some particular lamp model, and sometimes are made shorter, for attachment to other wires--often called the "wiring harness"--which then extend the remaining distance to the sockets. Representative patents exemplifying this standard configuration include U.S. Pat. No. 2,489,245 to Sola, U.S. Pat. No. 2,595,487 to Runge, U.S. Pat. No. 3,360,687 to Riesland, and U.S. Pat. No. 3,655,906 to Robb; as well as Canadian Patent 751,052 to Kukla.
Adherence to this basic form of ballast wiring has remained dominant in the industry despite issuance of many patents proposing seemingly reasonable variations. U.S. Pat. No. 2,487,468 issued in 1949 to Shirley R. Naysmith for one such variation--in which the wires from each end of the ballast terminate in respective half-connectors; these plug directly into mating half-connectors in lamp-socket assemblies, at the ends of the fixture respectively.
The Naysmith patent proposed that "all the wiring within the luminaire may be completed by merely plugging together the cable-carried receptacles to the fixed lamp holders." The inventor envisioned that fixture assembly would be thereby rendered so easy that "ballast units may be completed and pretested by the ballast manufacturer, the lamp holders by the lamp holder manufacturer, and shipped to the [installation] location in suitable lots without passing through the factory of the fixture manufacturer, thereby avoiding freight and handling, and the parts can be readily assembled on the job . . . " Naysmith's device is not a "leadless" ballast.
In U.S. Pat. No. 3,514,590, M. David Shaeffer proposed (1970) a leadless ballast, made to plug into a printed-circuit board that would--with a single backing plate--replace both the casing and the wiring of a fluorescent-lamp fixture. The lamp sockets as well as the plug-in ballast were to be supported at the underside of the printed-circuit board. Shaeffer's objective was that the entire fixture be amenable to assembly quickly and without the use of tools.
U.S. Pat. No. 3,569,694 of Oscar L. Comer posited in 1968 that a ballast-can coverplate be extended longitudinally beyond one end wall of the can, and that an array of laterally oriented connector pins be fitted to a vertical bracket on the baseplate extension. Short wires passed to these pins through the nearby end of the ballast can; and the pins in turn mated with a complementary array of laterally oriented female contacts mounted to the casing of the fixture. This unit thus might be called "almost leadless".
The plug-in concept was carried to its logical extreme in U.S. Pat. No. 4,674,015 of Daniel R. Smith, which in 1987 taught that the entire ballast should be plugged bodily sideways into a large receptacle in the casing. In Smith's leadless design, contact tabs on the interior wall of the receptacle engage mating contact tabs on the side wall of the ballast can.
U.S. Pat. No. 4,729,740 issued in 1988 to Crowe et al., showing a small printed-circuit board within the ballast can--and supporting all the other components in the can. In particular the internal circuit board supported at each end of the assembly a respective electrical connector for attachment of the several individual leads of a wiring harness leading to each half (i.e., each end) of the fixture respectively. Crowe's ballast too is thus a leadless configuration.
From Crowe's drawings it appears that his invention is intended primarily for use as one of the previously discussed "electronic ballast" types. His text, however, by its general language seems to suggest that the invention has broader application to more-conventional "magnetic" ballasts as well.
At each end of the assembly, Crowe's connector fits against the end wall of the can--except where the connector protrudes through a window cut in the end wall--and is longitudinally stabilized by grooves in the connector that receive the cut side edges of the window. We refer to this kind of mounting, in which the connector edges define a groove that makes a sliding engagement with the edges of a window in the end wall, as a "picture frame" mounting.
The firm with which we are associated, MagneTek Universal of Paterson, N.J., has introduced a leadless electronic ballast under the trademark "LUMINOPTICS" and covered by U.S. Pat. No. 4,277,728. It has a full-length circuit board generally analogous to Crowe's--but mounted to a flat plate that becomes the cover, rather than to the U-shaped body. It also has a second board that is much shorter and mounted vertically to the full-length board.
The LUMINOPTICS ballast is not potted, although some of the components are individually dipped. It has various modern features including a connection for computerized control, and a manual dimmer control.
A poke-in eight-contact wiring connector is provided at each end of the ballast, respectively. Each connector is mounted to a corresponding end of the full-length circuit board, accessible through a port in the associated end wall.
A groove defined in each of these connectors engages an inset flange formed at the bottom of the port, to stabilize the connector to the U-shaped body. A separate two-pin standard connector is installed in one end wall for power input.
Another leadless ballast design that uses an internal connector is disclosed by Burton et al. in U.S. Pat. No. 4,916,363 (1990), assigned to Valmont Industries, Inc. of Nebraska. Here the internal connector receives the wiring-harness wires either individually or in a connector-like carrier that organizes the wires into an array, but the internal connector is not mounted in the picture-frame style as in Crowe--and in fact is not in an end wall of the can at all.
Instead the internal connector is mounted in a transverse slot that extends all the way across the width of the bottom of the can, about a quarter or a third of the distance along the can from one end. At the side of the internal connector which faces toward that nearer end, the bottom of the can is formed in a shallow bevel that makes the connector face accessible for insertion of the wires.
The ballast can of Burton et al. is also formed with an inset longitudinal ledge (or, more strictly speaking, upside-down ledge) along each of its lower longitudinal corners. Each ledge is used for routing of wires from the connector in both longitudinal directions to the lamp sockets, and at each end is provided with "clamp portions"--apparently formed integrally with the ballast can--adapted to be bent over toward the inset ledge, to keep the wires on the ledge.
Because of the ledges along each lower corner, the cross-section of the can has a step at each corner. On one side of the transverse slot, the connector-surfaces abut or fit against inside surfaces of the can all the way down both sides and across the bottom, including the corner steps. Therefore the connector too is notched or stepped at its lower corners.
At the other side of the transverse slot in the can, a flat surface of the connector abuts the cut-off edge of the slot. As will be seen, these several surfaces abutments at three different orientations pose at least a challenge to attainment of effective seals during potting.
Another modern development in leadless ballasts, apparently not now the subject of an issued patent, is the line of ballasts available from the Valmont Electric Company (a subsidiary of Valmont Industries) under the commercial designation "XL Series". An XL ballast has a single half-connector mounted in one end wall of the ballast can, and formed as a receptacle.
That wall-mounted receptacle receives another half-connector, configured as a jack, which terminates the wiring harness. The receptacle fits within, and protrudes slightly through, a window cut in the end wall of the can; while a flange around the receptacle is provided to press against the inner surface of the end wall, all around the window.
In the Valmont XL Series ballasts the receptacle carries a row of male contact pins, which are the tips of rectangular-cross-section metal strips leading from an intermediate terminal block. The terminal block is positioned about an inch inside the can, and is apparently held generally suspended (before potting) in that region by electrical leads soldered to contacts on the electrical components.
In the XL Series configuration, during potting, two small ratchet-style locking tabs--one at each end of the half connector, respectively--hold the receptacle flange against the inside of the wall. These tapered snap tabs, based on our own testing of such fasteners, give a better seal than the picture-frame retainers discussed earlier--but here too, at a production-engineering stage prove overly sensitive to the possibility of tolerances adding up adversely.
Since the contacts in the receptacle are male, the jack of course carries female contacts; within the jack the female contacts are permanently secured to the ends of the wires in the harness. These wires leave the jack body through a surface that faces the end wall of the can, so that at least those wires which lead to lamp sockets at the same end of the fixture as the jack are bent in a tight "U" shape.
Of the several variants discussed above, only the last three seem to have become commercially important. The concept of a leadless ballast does seem to be gaining some ground in the fluorescent-lighting industry. In fact a significant effort has been mounted by Valmont Industries to declare such ballasts--and, mope particularly, the connector and pin configurations of the XL Series--an industry standard.
Perhaps the fluorescent-lighting industry could benefit from ballast standardization, but there is no standard yet. We believe that all of the above-discussed variations, including the two Valmont configurations, have important limitations which should be addressed and resolved before settling upon any of them, or even any combination of their features.
A few of the known features discussed above--especially the circuit-board mounting used in Crowe and the LUMINOPTICS ballast--appear adequate for some electronic ballasts, which are lighter and produce less vibration. As will be seen, however, such mounting is problematic for other electronic ballasts that do have relatively heavy radio-frequency-interference and power-factor filters, and also for the more-familiar magnetic ballasts, which still constitute by far the greatest fraction of ballast sales.
All or most of the remaining limitations seem to flow from inadequate recognition of several major characteristics of the overall process of ballast and fixture manufacturing, distribution, use and replacement. For specific reference we shall state these characteristics in the form of eight numbered "ground rules" for ballast design:
(1) The fluorescent-lighting industry is price competitive to an extreme. Profit margins in ballasts are correspondingly small, and production volumes are very high--so that manufacturing-cost advantages of only a fraction of a penny per ballast are likely to be significant.
(2) A major factor in ballast manufacturing cost is labor, particularly hand labor. Seconds lost in fussing with assembly or with touchy alignments and the like prior to potting, or later in wiping spilled or leaked potting potting material from the outside of each ballast, translate into major cost components.
(3) Material costs of course are also important, and militate strongly against use of additional intermediate components to perform limited functions. For example, the relatively expensive floating intermediate terminal block in the XL Series ballasts apparently is used primarily to obtain effective strain relief of the electrical leads inside the ballast can.
(4) Another cost-related consideration is that a ballast connector should be as compatible as practical with already-existing ballast-design and ballast-manufacturing techniques. Some changes in assembly-line equipment and layout or sequence can be very expensive, and as amortized--even over many hundreds of thousands of ballasts--can thereby add significantly to unit cost.
(5) Commendable wishes for industry standardization are not the same thing as actual achieved standardization. Any ballast configuration that is offered as a standard must offer users, distributors, fixture manufacturers and ballast manufacturers alike some reasonable means of coping with a protracted period of time during which standardization among manufacturers is incomplete. In addition, regardless of leadless-ballast standardization, it seems unlikely that the industry will achieve complete standardization of fixture lengths, or accordingly of wiring-harness lengths.
(6) Any proposed standard ballast must also accommodate effectively an even more protracted replacement or retrofit period. During such a period the new-style ballasts must be used to replace millions of used ballasts of many different configurations--but primarily the long-time standard ones shown in, for example, the Sola, Runge, Riesland, Robb and Kukla patents mentioned earlier. Therefore a ballast connector should accommodate replacement or retrofit of earlier conventional ballasts that have protruding leads.
(7) Fluorescent-lamp fixtures intrinsically are roughly handled, knockabout items that must be designed to intrinsically withstand careless handling, and some degree of improper installation. Consumers do not treat fixtures or ballasts as if they were, for example, laboratory instruments or personal computers; therefore it is a mistake for designers to so treat them.
(8) Magnetic (and some electronic) ballasts themselves contain heavy components that can generate significant internal forces due to mechanical shock and vibration in shipping and handling. Once in operation they also generate heat and develop forcible vibrations, which often increase with use. Successful ballast designs therefore must avoid not only use of fragile elements, but also elements that when heated or vibrated can damage other nearby standard components (such as wiring).
Based upon these ground rules 1 through 8, we shall now comment upon the several ballast variants discussed above. We wish to make clear that all of these devices may serve (or may have served) reasonably well for their intended purposes; the comments that follow will simply show that there remains some opportunity for improvement.
The Naysmith design violates ground rules 1, 3, 8 and 8, as it requires a ballast with preattached cables, at least long enough to reach the lamp sockets, and it provides every new ballast with two relatively expensive half-connectors and cables. At the outset, Naysmith's proposed system would thus be prohibitively expensive, in modern terms.
Moreover, the connectors and cables of an older Naysmith ballast being replaced are discarded with the old unit, even though the old connectors and cables usually are in perfectly good condition. Worse yet, to use the ballast with an older standard fixture, the expensive connectors and cables must be cut off and discarded at the outset.
Even for use with various models of a single manufacturer the design is undesirable. The manufacturer must assemble, and then the distributor must stock, ballasts with several different cable lengths. If the distributor is out of stock for a unit with a short cable, the buyer must settle for a more expensive one with a long cable.
The Shaeffer design violates at least ground rules 7 and 8. During handling, installation or replacement the weight of the ballast is likely to be inadvertently struck against the very large, expensive printed-circuit board--incurring the risk of damage to the board. As is well known, such damage is likely to be partially or entirely concealed and is likely to cause an electrical fault of the worst sort--namely, an intermittent one.
If proposed as an industry standard, it would also violate ground rules 4 through 6. Here, however, as contrasted with the Naysmith situation already discussed, the difficulty of using Shaeffer's ballast configuration in a conventional fixture would be essentially prohibitive. It is clear that Shaeffer's teachings are not intended to have any compatibility with existing or present standard fixtures.
Thus, as he explains, the electrical connections of his ballast terminate in an array of small connector pins in the coverplate. For use with a standard wiring harness, these pins would require some sort of mating connector added to the wire ends--or perhaps a solder joint.
Shaeffer does not address these possibilities, for the apparent reason that the connector pins would interfere with mounting of his ballast in a conventional fixture anyway. Plainly, use of that ballast in such a fixture would require far more than use of Naysmith's--i.e., more than merely cutting off and discarding expensive but unused components.
The Comer configuration too would violate ground rules 4 through 6, although in degree of incompatibility with earlier fixtures it is perhaps intermediate between the Naysmith and Shaeffer designs. In Comer's unit, some wires do extend out of the can, perhaps three to five centimeters, to his laterally mounted connectors; thus cutting off and discarding the connectors might possibly permit connection by means of wire nuts or the like to the stub wiring.
As will be evident, however, making connections to such short wires is difficult or at least awkward and annoying. In the course of the process a growing cluster of wire nuts would develop in a small region adjacent to the end of the can, requiring progressively greater dexterity and care to make each successive connection. Even removal of the Comer connectors and their mounting bracket--if indeed that were feasible without damaging the coverplate--would make available very little additional room for the new connections.
In addition Comer's ballast violates ground rules 1 through 3. The additional metal usage for the coverplate extension and connector bracket, and the hand-mounted individual connectors, would probably make Comer's design economically unfeasible.
Daniel Smith's ballast violates ground rules 4 through 8, for generally the same reasons as Shaeffer's ballast. If anything, Smith's configuration is more problematic with respect to retrofit: his contact tabs appear probably even more resistant to adaptation for use in older fixtures than Shaeffer's pins.
The Crowe ballast is particularly interesting, since it is relatively similar in outward appearance to other modern designs (including the LUMINOPTICS unit). It is also interesting because Crowe's patent contains some important teachings which are followed by one other patented design, but which we regard as incorrect.
For most ballasts--more specifically, for magnetic ballasts and those relatively heavy electronic ballasts that have power-factor or radio-frequency-interference filters--the Crowe configuration violates ground rules 7 and 8. During shipping and handling, the weight of the ballast components is likely to crack the internal circuit boards, causing damage even more obscure than that discussed above with respect to Shaeffer's large external circuit board. Crowe's circuit board is even more subject to damage due to vibration.
Whether caused by handling damage or vibration, damage to the circuit board in a Crowe ballast is even more likely to be intermittent. His circuit board is more directly coupled to heat developed within the electrical components of the ballast, and therefore more likely to flex during warmup. Flexure might not occur, however, until heat accumulates to nearly a steady-state operating condition, perhaps an hour after the lamp starts.
We believe that Crowe's invention also violates ground rules 1 and 2, at least for fully potted ballasts. We have experimented with connectors mounted by a "window frame" kind of mounting, of the general sort employed in Crowe's ballast, and found such mounting unacceptable. Problems with such mounts arise from the generally rough-work nature of the inexpensive sheet-metal forming procedures used in making ballast cans.
More specifically, we learned that the sometimes rough sheet-metal edges, and sometimes very substantial curvature of the metal, produced a much higher need for installation force than anticipated. When the window-frame grooves along the connector edge were widened to alleviate this problem in some units, then the fit was rendered loose or sloppy for other units that happened to be smoother or less curved.
Hence, if a window-frame mount is chosen to be relatively tight, extra assembly time and cost will often be required to force the connector into place--with caution needed to avoid slips that could cut the workers' hands on the metal edges. These operations could be particularly difficult in a ballast with a circuit board attached to each connector, as in Crowe.
On the other hand, if the mount is chosen to be relatively loose, then extra time and cost will often be required to wipe away the potting material that leaks around the edges of the connector in a loose mounting. In especially loose installations, our connectors actually floated upward in the potting material, as that material was poured, leading to what might be called "catastrophic leaks".
Thus, in summary, fit is critical in window-frame mounting. Special precautions of course could be taken to hold the connector in place, and perhaps also to press it firmly against the wall during initial stages of pouring the potting material; but these precautions would be unacceptably costly in terms of labor.
In Crowe's configuration the connector cannot float out of place because it is secured to the circuit board; but we regard circuit boards as undesirable in most ballasts, for the reasons already discussed. Thus as noted above we consider picture-frame mounting to violate ground rules 1 and 2.
Crowe provides connectors that receive discrete leads from the wiring harness individually, rather than grouped leads held in a half connector as in Burton and in the Valmont XL Series. Crowe explains:
"One . . . manufacturer has included an electrical connector . . . for interconnection thereto by a mating electrical connector. The disadvantage to having an electrical connector at the end of the discrete wires is that typically the fluorescent fixtures are not sold with a mating electrical connector. Therefore, the manufacturer of the ballast has to include both connector halves which increases the cost of the electrical ballast. Furthermore, the installer . . . must not only replace the ballast but must also terminate the discrete wires of the lighting to the mating half of the electrical connector. When replacing the ballast, the user . . . must buy a ballast which also carries an electrical connector which is matable with the electrical connector of the first ballast installed."
For several reasons, we believe that Crowe is incorrect in this teaching. First, he fails to recognize the two enormous benefits of using an external connector, whether prewired by a fixture manufacturer or attached later by an installer of a replacement ballast:
(1) After the external connector has once been permanently installed on the wiring harness and the harness tested, all ballast installations thereafter (including both the initial installation and all replacements) are far easier and simpler.
(2) More importantly, after the first test of the combined connector and harness, all later ballast installations are also rendered virtually foolproof with respect to correct wire-to-pin correspondence.
This latter point is most crucial, since the time required to make individual-lead connections is not merely the time required to plug in a single connector multiplied by the number of leads; to the contrary, great care (entailing extra time) must be taken to ensure that each lead is being connected to the proper contact.
Secondly, Crowe overlooks the fact that for new fixtures--when the ballast is sold on an OEM basis to the fixture manufacturer--that manufacturer will be willing to pay for the slight additional cost of the external half connector (partly offset by a small saving in labor cost for wiring and testing), in order to obtain the competitive advantage of being able to advertise especially easy ballast replacement.
Thirdly, turning now to use of a new-style leadless ballast for field replacement of older-style ballasts: there is a fallacy behind Crowe's assertion that the user must buy a replacement ballast that "also carries an electrical connector which is matable with the electrical connector of the first ballast installed."
What Crowe overlooks here is that, when a ballast meeting all the above-mentioned ground rules is introduced to the fluorescent-lighting industry, there may be greater reason to expect standardization of pin assignments and connector configurations. Thereafter all new ballasts would carry compatible connectors; Crowe's objections would then all die within one generation of ballasts.
Fourthly, also regarding new leadless ballasts used as field replacements, Crowe overlooks various possibilities for distributing the external half connector for use in field replacement. At first, of course, for a period of perhaps four to seven years virtually every leadless ballast sold for field-replacement use would require such an external half connector; therefore during that preliminary transitional period it would be simplest to include one external half connector (and its price) with every new replacement ballast.
After that, manufacturers could make an external connector available to retailers for distribution separately as an "adapter", either at a nominal price or free upon request. These procedures, if judiciously timed, would limit the manufacturer's added cost to, on average, a small fraction of the cost of one external half connector for each older-style ballast that is replaced.
Fifthly, and still as to field replacements, Crowe overlooks the possibility that to "terminate the discrete wires . . . to the mating half" the installer need not necessarily do any mope work than would be required to make individual connections to Crowe's internal connector! That is, the wiring provisions in the external half connector may be made of the poke-in-and-lock type.
Stripped discrete leads would then be simply inserted into the rear of the external half connector, just as is the case with Crowe's connector. The poke-in connections would be substantially permanent, but release cams could be included in the half connector for prompt correction of wiring errors.
Sixthly, Crowe fails to realize that providing for use of an external half connector is not necessarily the same thing as requiring one. That is, allowing for use of an external half connector can be made compatible with attachment of the wiring harness discrete leads to the can-mounted half connector individually.
In other words, the benefits of using an external half connector may be achieved while retaining the user's options to wire replacement ballasts without one. Parts of this strategy are shown by Burton, whose ballast design we shall discuss next.
Burton's ballast violates ground rules 1 and 2, because the geometry of the connector and of its centralized mounting is inherently subject to leakage. The reason for this vulnerabilility is that the can and the connector both have steps at their two lower corners.
At each step there is one horizontal segment and one vertical segment. In addition there is a third horizontal segment across the floor of the can.
If the tolerance of all five of these segment lengths, as established in the sheet-metal forming steps, is not held to perhaps 3/4 millimeter (0.03 inch) or better, potting-material leakage is likely to be substantial. Ballast-can construction, however, for the necessary economies desired according to ground rule 2, is inherently of a coarse character; fine tolerances are rather beyond the norm--at least for a multisegment shape as required by the Burton geometry.
This is particularly so if one takes into consideration the great variation of bending properties and resilience in different material lots. Even apart from varying impurity content and the like, normal cold-rolled steel used in ballast cans is typically 0.66.+-.0.08 mm (0.026.+-.0.003 inch) in thickness: that tolerance of nearly twelve percent of course generates large variations in strength, resilience, etc.
Either inordinate labor cost must be incurred to hold unusually tight sheet-metal forming tolerances to avoid leakage, or extra labor must be expended in wiping away potting material after pouring. In either event, the Burton configuration also demands extremely careful positioning (or some other sealing technique) to avoid leakage at the abutment between the vertical face of the connector and the straight cut edge along the beveled-floor segment of the can.
The Burton ballast also violates ground rule 8, in Burton's provisions for routing wires of the harness from the centrally mounted connector in both directions along the ballast to the lamp sockets. Concededly, Burton's previously described ledges and cable clamps do impose some orderliness upon the wire runs.
Presumably this is an effort to avoid damage by pinching of stray leads between the ballast housing and the fixture casing. Burton's solution, however, appears to be counter-productive.
To the extent that the character of the clamps can be determined from the Burton patent, they appear to be metallic, and in fact unitary with the other portions of the ballast can. It would seem that using such clamps, likely with sharp edges, to secure wires along the ballast-can ledge actually creates a risk of damage to the wires or their insulation. The significance of such damage will be apparent.
Forming the clamps over the wires also represents an undesirable additional manufacturing cost--a violation of ground rules 1 and 3. Furthermore, the clamps make installation or replacement much more difficult.
Thus Burton's ballast violates ground rules 1 through 3, and 8. It does demonstrate, however--as mentioned earlier--that a ballast connector may be configured to receive wiring-harness leads either (a) as a group held in a connector, or (b) individually if the connector is unavailable.
Burton's wiring-harness carrier 66 serves virtually as a connector body, to hold the individual wires together in a standardized array that matches the contact array of the mating connector in the ballast. The system therefore provides both quick connection and the essential certainty of correct wiring, and so takes a step in the right direction with respect to ground rules 5 and 6.
The individual bare-wire ends held by Burton's carrier directly engage poke-in contacts of the connector that is mounted in the ballast. Therefore a person who does not have Burton's carrier can nevertheless insert the bared ends of individual or discrete wires directly into the same poke-in contacts, to attach an older-style fixture (which has no wire carrier) to the ballast.
Of course this is not as convenient as using an external carrier or connector body, but is as convenient as any other system for attaching wires individually--i.e., as convenient as earlier conventional systems using wire nuts, or using poke-in systems such as Crowe's. Hence Burton's connection system facilitates field replacement of old-style ballasts, as well as OEM installation.
Burton's apparatus shows that the benefit of an external half connector may be kept while retaining the user's option to wire replacement ballasts without one. As Burton's patent fails to mention or even suggest this dual function, however, it is not clear whether Burton obtained this benefit intentionally or inadvertently; furthermore, the specific mechanics of his system are questionable on several grounds, as follows.
Burton's system uses poke-in contacts in the ballast-mounted connector. These poke-in wiring connections between the ballast and the wiring harness constitute the entire mechanical system for holding the harness to the connector.
That is, the wiring system is required to serve as its own strain-relief system. We consider such a confusion between the functions of electrical contact and mechanical integrity to be relatively undesirable industrial practice, implicating indirectly ground rule 8 above.
If excessive withdrawal force is applied to the wires while they are restrained by the poke-in contacts, the tangs inside the poke-in connector may damage the wire ends--either jamming them within the poke-in cavities, or weakening them so that they fail later under vibration, or possibly deforming them so that they cannot later make good contact with the poke-in contacts of another ballast.
Burton provides a "release comb" to disengage all the poke-in contacts at once, to allow for removal of the external wires with their attached carrier. This release comb is relatively wide and short, and therefore appears susceptible to cocking and then binding in its guides, particularly if a user attempts to operate it after the ballast has been in operation under typical conditions of heat, accumulating dirt, and vibration for several years.
Burton's patent does not state whether the comb is stowed permanently in its guides ready for use in field replacement, or is to be kept nearby for such use. (If the former, the assembly sequencing must be selected to avoid potting the comb; and if the latter, the comb is likely to be lost before it can be used.) Whichever may be the situation, the user must first find the comb and otherwise see to its proper positioning--partially concealed above the wiring carrier.
The user must then try to slide the comb longitudinally, relative to the housing, in a short operating recess adjacent to the ballast-mounted connector: the release comb operates in cramped quarters at best.
Most drawbacks of Burton's ballast arise at least partly from the centralized location of the connector. We therefore submit that such centralized mounting is undesirable.
As has been shown in discussion of the Crowe ballast, however, problems also arise in prior-art efforts to mount a connector at an end (or at each end) of the can. This assertion is validated by consideration of the XL Series ballast, with its end-mounted connector.
That ballast appears to violate ground rules 1 through presented above. We shall take these points in order.
Within the ballast can, the XL ballast apparently requires an additional, costly intermediate terminal block for strain relief, as well as custom-made and custom-assembled flat metal strips that serve as pins and intermediate connectors. Extra labor--which may appear partly as material cost, if the assembly is bought complete for OEM use--is also required to make connections at both sides of this terminal strip.
In potting, the XL ballast relies upon a pair of tapered or ratchet-type snaps to hold the connector flange against the inside of the end wall. This technique relies on controlled deformation of both the plastic snaps and the metal edges. Formed sheet metal, however, is subject to uncontrolled bending or warping, particularly near corners. Rolled and punched sheet-metal construction is inherently coarse.
Under these conditions, in our experience, the window will sometimes seem too wide to yield a reliable seal, and sometimes too narrow for the snaps to pass through, with a reasonable amount of force. In either event, the result is additional labor, extra attention for seconds or minutes--to either force the snaps in, or wipe away potting-material leakage later. Tolerances can be controlled to avoid these problems, but the cost of doing so is then objectionable.
The XL unit also uses additional current-carrying components, at least within the ballast housing. This too increases cost without clear advantage.
As to ground-rule 4, the extra terminal strip in the XL system also requires an additional assembly step, rendering the unit relatively incompatible with a standard assembly line. In addition the extra connection introduces undesirable electrical resistance, which can be significant especially in some so-called "rapid start" filament circuits that operate on as little as three volts.
Outside the can, the XL Series ballast fails to answer the challenge posed by Crowe: connection is possible only by means of the external half connector, with no mitigating provision for field replacement. The external half connector does not appear to be of an easy-to-wire (e.g., poke-in) type such as we have described above; and there is no suggestion in the XL Series literature of any arrangement for making the external connectors available to users separately for field replacement.
In addition, the previously mentioned reverse wire dress of the external connector can only serve as an invitation to damage during shipping, handling, or field replacement. With that we reach ground rule 7.
In view of all the foregoing it appears clear that the prior art has not yielded a fluorescent-lamp leadless ballast, or leadless-ballast-and-harness combination as appropriate to the context, that makes use of an external half connector for its very important benefits while satisfying all of ground rules 1 through 8. A long-felt need of the fluorescent-lighting industry--and of the users of fluorescent lighting--has thus gone unmet.